dl_datasets() and e_t_datasets() should be more error-tolerant

[jeanpaulrsoucy]

Errors should not stop/reset the entire update process.

[jeanpaulrsoucy]

This has been accomplished, for the most part.

[jeanpaulrsoucy]

Function-stopping error when trying to retry multiple datasets at once (retrying one dataset seems to work):

WAITING 10 SECONDS BEFORE RETRY...
RETRY SUCCESSFUL:  2e7a5549-92ae-473d-a97a-7b8e0c1ddbbc e00e2148-b0ea-458b-9f00-3533e0c5ae8e
Error in ds[[ds_failed]] <- tryCatch({ : 
  more elements supplied than there are to replace

It seems to be trying to do both datasets at once.

[jeanpaulrsoucy]

If reading from Google Sheets fails, the script ends. Instead, the section should be skipped (and the error reported).

[jeanpaulrsoucy]

Additional processing in e_t_datasets should also be more error tolerant: if the main function returns NA (e.g., because the dataset was not available), further processing (e.g., process_hr_names, dplyr::summarize, etc.) should either be skipped or done with error catching (better, since the functions could potentially fail for other reasons).

Update: edde832a81128ee041ef8c29313c5827d0ab2fbd fixed the above issue for process_funs but not for one-off use of dplyr functions.

[jeanpaulrsoucy]

Values that take the maximum value of two datasets (i.e., provincial and federal vaccine distribution numbers) cause an error in the script when one of the values is not available. This should be rectified by writing an explicit function with error tolerance.

[jeanpaulrsoucy]

The sheets_merge function and other functions pulling/pushing to Google Sheets should have error-tolerance added, since these are a common point of failure.